Single-cell spatial transcriptomics reveals a dystrophic trajectory following a developmental bifurcation of myoblast cell fates in facioscapulohumeral muscular dystrophy.

Facioscapulohumeral muscular dystrophy (FSHD) is linked to abnormal derepression of the transcription activator DUX4. This effect is localized to a low percentage of cells, requiring single-cell analysis. However, single-cell/nucleus RNA-seq cannot fully capture the transcriptome of multinucleated large myotubes. To circumvent these issues, we use multiplexed error-robust fluorescent in situ hybridization (MERFISH) spatial transcriptomics that allows profiling of RNA transcripts at a subcellular resolution. We simultaneously examined spatial distributions of 140 genes, including 24 direct DUX4 targets, in in vitro differentiated myotubes and unfused mononuclear cells (MNCs) of control, isogenic D4Z4 contraction mutant and FSHD patient samples, as well as the individual nuclei within them. We find myocyte nuclei segregate into two clusters defined by the expression of DUX4 target genes, which is exclusively found in patient/mutant nuclei, whereas MNCs cluster based on developmental states. Patient/mutant myotubes are found in "FSHD-hi" and "FSHD-lo" states with the former signified by high DUX4 target expression and decreased muscle gene expression. Pseudotime analyses reveal a clear bifurcation of myoblast differentiation into control and FSHD-hi myotube branches, with variable numbers of DUX4 target-expressing nuclei found in multinucleated FSHD-hi myotubes. Gene coexpression modules related to extracellular matrix and stress gene ontologies are significantly altered in patient/mutant myotubes compared with the control. We also identify distinct subpathways within the DUX4 gene network that may differentially contribute to the disease transcriptomic phenotype. Taken together, our MERFISH-based study provides effective gene network profiling of multinucleated cells and identifies FSHD-induced transcriptomic alterations during myoblast differentiation.

New rabies viral resources for multi-scale neural circuit mapping.

Comparisons and linkage between multiple imaging scales are essential for neural circuit connectomics. Here, we report 20 new recombinant rabies virus (RV) vectors that we have developed for multi-scale and multi-modal neural circuit mapping tools. Our new RV tools for mesoscale imaging express a range of improved fluorescent proteins. Further refinements target specific neuronal subcellular locations of interest. We demonstrate the discovery power of these new tools including the detection of detailed microstructural changes of rabies-labeled neurons in aging and Alzheimer's disease mouse models, live imaging of neuronal activities using calcium indicators, and automated measurement of infected neurons. RVs that encode GFP and ferritin as electron microscopy (EM) and fluorescence microscopy reporters are used for dual EM and mesoscale imaging. These new viral variants significantly expand the scale and power of rabies virus-mediated neural labeling and circuit mapping across multiple imaging scales in health and disease.

Measuring the composition of the tumor microenvironment with transcriptome analysis: past, present and future.

Interactions between tumor cells and immune cells in the tumor microenvironment (TME) play a vital role the mechanisms of immune evasion, by which cancer cells escape immune elimination. Thus, the characterization and quantification of different components in the TME is a hot topic in molecular biology and drug discovery. Since the development of transcriptome sequencing in bulk tissue, single cells and spatial dimensions, there are increasing methods emerging to deconvolute and subtype the TME. This review discusses and compares such computational strategies and downstream subtyping analyses. Integrative analyses of the transcriptome with other data, such as epigenetics and T-cell receptor sequencing, are needed to obtain comprehensive knowledge of the dynamic TME.

[Box: see text].

A Case of Complex Giant Rhinophyma Treated With Cosmetic Surgery.

ABSTRACT: Rhinophyma leads to severe facial deformities and significant social pressure for patients. Patients often seek medical intervention due to cosmetic defects and functional impairments, such as nasal congestion and airway collapse. Currently, there are numerous treatment modalities for rhinophyma, each with distinct advantages and disadvantages, leading to a lack of consensus in nasal vegetation management. Severe thickening in the nasal area can obstruct breathing through external nasal valve blockage, necessitating appropriate management for relief. This article presents a case study involving severe rhinophyma with respiratory obstruction that was successfully treated using incomplete resection followed by reconstruction to restore normal nasal contour. This not only achieved an upright position for nasal columella but also improved nasal contour to achieve normal appearance levels while completely relieving respiratory tract obstruction and enhancing patients' ventilation function. This method is easily performed without requiring additional expensive equipment, making it economically feasible even in ordinary medical centers while enabling patients to achieve a high quality of life.

Degenerate mapping of environmental location presages deficits in object-location encoding and memory in the 5xFAD mouse model for Alzheimer's disease.

A key challenge in developing diagnosis and treatments for Alzheimer's disease (AD) is to detect abnormal network activity at as early a stage as possible. To date, behavioral and neurophysiological investigations in AD model mice have yet to conduct a longitudinal assessment of cellular pathology, memory deficits, and neurophysiological correlates of neuronal activity. We therefore examined the temporal relationships between pathology, neuronal activities and spatial representation of environments, as well as object location memory deficits across multiple stages of development in the 5xFAD mice model and compared these results to those observed in wild-type mice. We performed longitudinal in vivo calcium imaging with miniscope on hippocampal CA1 neurons in behaving mice. We find that 5xFAD mice show amyloid plaque accumulation, depressed neuronal calcium activity during immobile states, and degenerate and unreliable hippocampal neuron spatial tuning to environmental location at early stages by 4 months of age while their object location memory (OLM) is comparable to WT mice. By 8 months of age, 5xFAD mice show deficits of OLM, which are accompanied by progressive degradation of spatial encoding and, eventually, impaired CA1 neural tuning to object-location pairings. Furthermore, depressed neuronal activity and unreliable spatial encoding at early stage are correlated with impaired performance in OLM at 8-month-old. Our results indicate the close connection between impaired hippocampal tuning to object-location and the presence of OLM deficits. The results also highlight that depressed baseline firing rates in hippocampal neurons during immobile states and unreliable spatial representation precede object memory deficits and predict memory deficits at older age, suggesting potential early opportunities for AD detecting.

An individualized causal framework for learning intercellular communication networks that define microenvironments of individual tumors.

Cells within a tumor microenvironment (TME) dynamically communicate and influence each other's cellular states through an intercellular communication network (ICN). In cancers, intercellular communications underlie immune evasion mechanisms of individual tumors. We developed an individualized causal analysis framework for discovering tumor specific ICNs. Using head and neck squamous cell carcinoma (HNSCC) tumors as a testbed, we first mined single-cell RNA-sequencing data to discover gene expression modules (GEMs) that reflect the states of transcriptomic processes within tumor and stromal single cells. By deconvoluting bulk transcriptomes of HNSCC tumors profiled by The Cancer Genome Atlas (TCGA), we estimated the activation states of these transcriptomic processes in individual tumors. Finally, we applied individualized causal network learning to discover an ICN within each tumor. Our results show that cellular states of cells in TMEs are coordinated through ICNs that enable multi-way communications among epithelial, fibroblast, endothelial, and immune cells. Further analyses of individual ICNs revealed structural patterns that were shared across subsets of tumors, leading to the discovery of 4 different subtypes of networks that underlie disparate TMEs of HNSCC. Patients with distinct TMEs exhibited significantly different clinical outcomes. Our results show that the capability of estimating individual ICNs reveals heterogeneity of ICNs and sheds light on the importance of intercellular communication in impacting disease development and progression.

Spatial coding defects of hippocampal neural ensemble calcium activities in the triple-transgenic Alzheimer's disease mouse model.

Alzheimer's disease (AD) causes progressive age-related defects in memory and cognitive function and has emerged as a major health and socio-economic concern in the US and worldwide. To develop effective therapeutic treatments for AD, we need to better understand the neural mechanisms by which AD causes memory loss and cognitive deficits. Here we examine large-scale hippocampal neural population calcium activities imaged at single cell resolution in a triple-transgenic Alzheimer's disease mouse model (3xTg-AD) that presents both amyloid plaque and neurofibrillary pathological features along with age-related behavioral defects. To measure encoding of environmental location in hippocampal neural ensembles in the 3xTg-AD mice in vivo, we performed GCaMP6-based calcium imaging using head-mounted, miniature fluorescent microscopes ("miniscopes") on freely moving animals. We compared hippocampal CA1 excitatory neural ensemble activities during open-field exploration and track-based route-running behaviors in age-matched AD and control mice at young (3-6.5 months old) and old (18-21 months old) ages. During open-field exploration, 3xTg-AD CA1 excitatory cells display significantly higher calcium activity rates compared with Non-Tg controls for both the young and old age groups, suggesting that in vivo enhanced neuronal calcium ensemble activity is a disease feature. CA1 neuronal populations of 3xTg-AD mice show lower spatial information scores compared with control mice. The spatial firing of CA1 neurons of old 3xTg-AD mice also displays higher sparsity and spatial coherence, indicating less place specificity for spatial representation. We find locomotor speed significantly modulates the amplitude of hippocampal neural calcium ensemble activities to a greater extent in 3xTg-AD mice during open field exploration. Our data offer new and comprehensive information about age-dependent neural circuit activity changes in this important AD mouse model and provide strong evidence that spatial coding defects in the neuronal population activities are associated with AD pathology and AD-related memory behavioral deficits.

Validation of a Chin Retrusion Scale for Chinese Subjects.

ABSTRACT: Chin augmentation procedures are gaining in popularity. The purpose of this study was to validate the China (Allergan) Chin Retrusion Scale (CACRS) and to evaluate the reliability of the scale. A team of 10 physicians based in Beijing, China, consisting of 1 principal investigator and 9 independent raters, assessed a pool of subject images. Using standardized equipment to capture 2-dimensional images, 100 subjects were photographed from a left, lateral, 90° view. Two measures were used to assess the validity of the scale: rater evaluations of clinically significant differences in 52 pairs of images and rater assessment of chin retrusion using the CACRS to grade the 100 images. The CACRS demonstrated almost perfect inter-rater agreement during 2 validation sessions, with intraclass correlation coefficients of 0.89 and 0.90 at sessions 1 and 2, respectively. Overall inter-rater agreement for all 9 raters was almost perfect, with a weighted kappa of 0.82. Based on inter- and intra-rater agreement, the CACRS was validated for physician ratings of chin retrusion in Chinese subjects. Results from clinical significance evaluations indicated that a 1-grade difference between image pairs was considered clinically meaningful for the CACRS. The CACRS is a validated and reliable photonumeric scale for the assessment of chin retrusion in Chinese men and women. The scale is a clinically significant reference tool for evaluating the effect of chin augmentation.

Additional adjuvant capecitabine in early breast cancer patients: a meta-analysis of randomized controlled trials.

Aims: To assess the efficacy and safety of adjuvant capecitabine in early breast cancer patients. Methods: A literature search of databases was conducted to identify randomized controlled trials reporting the efficacy and toxicity of capecitabine as adjuvant therapy in early breast cancer patients. Results: Six studies were eligible and included a total of 6941 patients. Disease-free survival (hazard ratio = 0.79; 95% CI = 0.71-0.88; p < 0.0001) was significantly improved with additional capecitabine, whereas improvement in overall survival (OS) was not significant. The more pronounced benefits in both disease-free survival and OS were observed among triple-negative breast cancer patients. Conclusion: Additional capecitabine in the adjuvant setting conferred substantial disease-free survival benefit and a tendency toward improved OS. Triple-negative breast cancer patients can benefit from capecitabine irrespective of the administration sequence. Capecitabine may be considered a preferred additional treatment for early-stage triple-negative breast cancer patients, and sequential capecitabine can serve as an alternative choice for patients with poor tolerance.

Lay abstract The authors' meta-analysis focused on the adjuvant role of capecitabine in early-stage breast cancer patients. The authors combined data from different studies to show that disease-free survival was significantly improved with additional capecitabine as adjuvant chemotherapy. The more pronounced survival benefits were observed among triple-negative breast cancer patients irrespective of the administration sequence (concurrent/sequential). Capecitabine may be considered a preferred additional treatment for early-stage triple-negative breast cancer patients, and sequential capecitabine can serve as an alternative choice for patients with poor tolerance.

The factors influencing the unmet effects of alpha hydroxy acid peels on acne vulgaris: A single tertiary center experience.

Acne vulgaris is a common condition. Alpha hydroxy acid (AHA) peels have been previously recommended as an option for patients with intolerance to first-line treatments especially when long-term use. The safety and efficacy of AHA peels for acne have been established, but the factors influencing the ultimate effects are unknown. We recruited patients diagnosed with acne who were intolerant to or refused the first-line treatments from July 2017 to December 2019 at our hospital and retrospectively collected the medical and demographic information of patients treated with a full course of AHA peels; data collected included age, sex, treatment history, compliance status, and efficacy after treatment. The efficacy score was defined by revised scales: 2 points indicated significant improvement, 1 point indicated mild improvement, and 0 point indicated no improvement. Additionally, only efficacy scores of 2 points were classified in the satisfactory group; the others were classified in the unmet effect group. Analyses were used to evaluate the potential influencing factor(s). A total of 141 patients (120 females, 21 male) were included in the final analysis. The patients in the satisfactory group were significantly older (higher proportion aged >28 years) (47.4% vs 29.5%, P = .046) and showed better compliance (88.7% vs 54.5%, P = .001) than the patients in the unmet effect group. Multiple analyses confirmed the effects of younger age (OR 2.70, 95% CI 1.15-6.34, P = 022) and poor compliance (OR 2.74, 95% CI 1.74-4.32, P = .001) on the unmet effect. Unsatisfactory compliance and age might decrease the effects of AHA peels on acne.

Retracted: Antiproliferative Effects of Matricine in Gemcitabine-Resistant Human Pancreatic Carcinoma Cells Are Mediated via Mitochondrial-Mediated Apoptosis, Inhibition of Cell Migration, Invasion Suppression, and Mammalian Target of Rapamycin (mTOR)-TOR/PI3K/AKT Signalling Pathway.

An editorial decision has been made to retract this manuscript due to breach of publishing guidelines, following the identification of non-original and manipulated figures. Reference: Kaifeng Fang, Li Wang, LuJia Chen, Tao Liu, Zhi Fang: Antiproliferative Effects of Matricine in Gemcitabine-Resistant Human Pancreatic Carcinoma Cells Are Mediated via Mitochondrial-Mediated Apoptosis, Inhibition of Cell Migration, Invasion Suppression, and Mammalian Target of Rapamycin (mTOR)-TOR/PI3K/AKT Signalling Pathway. Med Sci Monit 2019; 25:2943-2949. 10.12659/MSM.914244.

Systematic discovery of the functional impact of somatic genome alterations in individual tumors through tumor-specific causal inference.

Cancer is mainly caused by somatic genome alterations (SGAs). Precision oncology involves identifying and targeting tumor-specific aberrations resulting from causative SGAs. We developed a novel tumor-specific computational framework that finds the likely causative SGAs in an individual tumor and estimates their impact on oncogenic processes, which suggests the disease mechanisms that are acting in that tumor. This information can be used to guide precision oncology. We report a tumor-specific causal inference (TCI) framework, which estimates causative SGAs by modeling causal relationships between SGAs and molecular phenotypes (e.g., transcriptomic, proteomic, or metabolomic changes) within an individual tumor. We applied the TCI algorithm to tumors from The Cancer Genome Atlas (TCGA) and estimated for each tumor the SGAs that causally regulate the differentially expressed genes (DEGs) in that tumor. Overall, TCI identified 634 SGAs that are predicted to cause cancer-related DEGs in a significant number of tumors, including most of the previously known drivers and many novel candidate cancer drivers. The inferred causal relationships are statistically robust and biologically sensible, and multiple lines of experimental evidence support the predicted functional impact of both the well-known and the novel candidate drivers that are predicted by TCI. TCI provides a unified framework that integrates multiple types of SGAs and molecular phenotypes to estimate which genome perturbations are causally influencing one or more molecular/cellular phenotypes in an individual tumor. By identifying major candidate drivers and revealing their functional impact in an individual tumor, TCI sheds light on the disease mechanisms of that tumor, which can serve to advance our basic knowledge of cancer biology and to support precision oncology that provides tailored treatment of individual tumors.

Enhanced production of L-methionine in engineered Escherichia coli with efficient supply of one carbon unit.

OBJECTIVE: L-methionine is an important sulfur-containing amino acid essential for humans and animals. Its biosynthesis pathway is complex and highly regulated. This study aims to explore the bottleneck limiting the improvement of L-methionine productivity and apply efficient strategies to increase L-methionine production in engineered E. coli. RESULTS: The enzyme O-succinylhomoserine sulfhydrylase involved in thiolation of OSH to form homocysteine was overexpressed in the engineered strain E. coli W3110 IJAHFEBC/PAm, resulting in L-methionine production increased from 2.8 to 3.22 g/L in shake flask cultivation. By exogenous addition of L-glycine as the precursor of one carbon unit, the titer of L-methionine was increased to 3.68 g/L. The glycine cleavage system was further strengthened for the efficient one carbon unit supply and a L-methionine titer of 3.96 g/L was obtained, which was increased by 42% compared with that of the original strain. CONCLUSIONS: Insufficient supply of one carbon unit was found to be the issue limiting the improvement of L-methionine productivity and its up-regulation significantly promoted the L-methionine production in the engineered E. coli.

Morphological Classification of the Zygomatic Regions in Young Chinese Han Males and Females Based on a Simulated Moiré Pattern.

The complex curved contours of the zygomatic regions are difficult to analyze. Therefore, a better evaluation medium must be developed. We aimed to examine and summarize the morphological characteristics of the zygomatic region by using a moiré pattern map and computer algorithm. In this cross-sectional study, we collected three-dimensional images of the facial contours of 251 Han Chinese youth and established a morphological moiré map database. Clustering analysis using a computer algorithm was applied to obtain the zygomatic morphologies for classification. Aesthetic evaluation was performed to summarize the characteristics of the zygomatic types and provide reference for the preoperative morphological design of the midface. Zygomatic regions were morphologically classified into five types. Each type had its typical feature in the moiré pattern map. The moiré stripes in the left zygomatic region formed an "Ω" shape outward and downward in type 1, and they tended to be diagonal like "\\\" in type 2, smoothly curved like ")))" in type 3, vertical like "|||" in type 4, and diagonal like "///" in type 5. The aesthetic evaluation outcome indicated that the integrally flat zygoma (type 4) was more aesthetically pleasing among males, and the integrally prominent zygoma (type 3) was more aesthetically pleasing among females. Five morphological contour types of the zygoma were classified among the Chinese Han males and females based on the simulated moiré pattern. This morphological classification would aid in preparing a guide for clinical diagnosis and surgical planning.

Antiproliferative Effects of Matricine in Gemcitabine-Resistant Human Pancreatic Carcinoma Cells Are Mediated via Mitochondrial-Mediated Apoptosis, Inhibition of Cell Migration, Invasion Suppression, and Mammalian Target of Rapamycin (mTOR)-TOR/PI3K/AKT Signalling Pathway.

BACKGROUND Pancreatic cancer is a major cause of mortality worldwide. Inefficient drugs, their adverse effects, and the development of drug resistance make it difficult to curb the growing incidence of pancreatic cancer. Against this backdrop, the development new drug regimens with no or negligible adverse effects is imperative. We assessed the anticancer effects of a plant-derived sesquiterpene - matricine - against capan-2 pancreatic cancer cells. MATERIAL AND METHODS Cell viability was determined by MTT assay. AO/EB, DAPI, and annexin V/PI staining were used to detect apoptosis. Transwell assays were used for monitoring of cell migration and invasion. Immunoblotting was used to examine the expression of proteins. RESULTS The results showed that matricine halted the proliferation of capan-2 cells, with minimal toxic effects on normal pancreatic cells. The anticancer effects were due to the induction of apoptotic cell death, which was allied with activation of caspases 3 and 9, upregulation of Bax, and downregulation of Bcl-2. Moreover, matricine suppressed the migration and invasive abilities of pancreatic cancer cells at IC50. We also assessed the effects of matricine on the mTOR/PI3K/AKT signalling pathway. We found that matricine efficiently blocked this pathway, suggesting the anticancer potential of matricine. CONCLUSIONS Matricine induced antiproliferative effects in capan-2 human pancreatic cancer cells through inducing apoptosis, caspase activation, inhibition of cell migration and invasion, and blocking the mTOR/PI3K/AKT signalling pathway.

Short Nose Correction: Septal Cartilage Combined With Ethmoid Bone Graft.

BACKGROUND: Multiple methods are employed to correct short nose deformities, with septal extension graft representing the first choice for Asians. However, the volume of the septal cartilage in many Asian patients is not sufficient to be used alone for such operation. The present work developed a new method combining the septal cartilage with ethmoid bone graft to overcome this issue in Asian patients with short noses. METHODS: Thirty-five women with short noses underwent septal extension graft from February 2015 to March 2017. The endoscopic technique was utilized to harvest the ethmoid bone to enhance the L-strut structure. An L-strut, comprising 0.8 cm segments of the caudal and dorsal cartilaginous septa, is left altered in order to harvest more cartilage for septal extension. The harvested the septal cartilage, approximately 1.0 mm thick and 16 to 20 mm long, underwent grafting on one side of the caudal septum. This was followed by alar cartilage fixation at the septal cartilage graft end. Finally, nose length, nasal tip projection and nasolabial angle were assessed before and after the surgery. RESULTS: Septal cartilage combined with ethmoid bone graft yielded an adequate nose lengthening and reduced nostril show, also in individuals showing extremely little septal cartilage. CONCLUSIONS: This new method is effective for short nose deformity correction in Asian patients. Combining the septal cartilage with ethmoid bone graft yields ideal outcome without overt complications, achieving commendable patient satisfaction.

A Novel Technique for Short Nose Correction in Asians: M-Shaped Conchal Cartilage Combining With Septal Extension Graft.

BACKGROUND: Multiple techniques are available for short nose deformity correction. The septal extension graft represents the commonest method employed in Asians. However, a large number of surgeons hardly obtain esthetically satisfactory results since the majority of designs do not reflect the normally encountered surface anatomy of the nasal tip cartilage. OBJECTIVE: The authors designed a novel technique, which combined the M-shaped conchal cartilage with the septal extension graft for overcoming the above shortcoming in Asians. METHODS: Between February 2013 and March 2016, 33 patients presenting short nose deformity were surgically treated with the M-shaped conchal cartilage combined with the septal extension graft. The graft was an altered septal extension graft using the septal cartilage alongside the conchal cartilage. The harvested septal cartilage was located to the caudal septum and fixed with sutures. The conchal cartilage was trimmed into 2 strips, which were sutured together in an M-shape and firmly fixed in a bilateral manner to the caudal septal extension graft. Then, the alar cartilage was fixed with the M-shape graft. In all patients, nasal lengths, nasal tip projections, and nasolabial angles were assessed before and after surgery, respectively. RESULTS: Nasal tip projections and nasal lengths showed remarkable increases, while columellar-labial angles were overtly decreased, in individuals surgically treated by the novel technique. CONCLUSIONS: This study presents a new method for correcting short nose deformity in Asians. The M-shaped conchal cartilage combined with the septal extension graft may effectively lengthen the nose while closely representing the actual surface of the esthetic nasal tip.

Lip Morphology and Aesthetics: Study Review and Prospects in Plastic Surgery.

The lip profile plays an important role in the perception of facial aesthetics; lip morphology and aesthetics research is receiving increasing attention. The advancement of research tools such as three-dimensional imaging technology has led to the clarification of lip morphologic and aesthetic characteristics. After studies of lip characteristics according to gender, ethnicity and age provided basic data, studies on lip aesthetics have been conducted by scholars worldwide. These studies could provide a basic theory to support diagnosis and treatment options, as well as the basis for evaluative criteria for precise treatment and technical improvements. According to the conclusions of the above studies, new ideas for cosmetic surgery design, including lip, perioral and labial-facial relationships, have been discovered.Level of Evidence V This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

Learning a hierarchical representation of the yeast transcriptomic machinery using an autoencoder model.

BACKGROUND: A living cell has a complex, hierarchically organized signaling system that encodes and assimilates diverse environmental and intracellular signals, and it further transmits signals that control cellular responses, including a tightly controlled transcriptional program. An important and yet challenging task in systems biology is to reconstruct cellular signaling system in a data-driven manner. In this study, we investigate the utility of deep hierarchical neural networks in learning and representing the hierarchical organization of yeast transcriptomic machinery. RESULTS: We have designed a sparse autoencoder model consisting of a layer of observed variables and four layers of hidden variables. We applied the model to over a thousand of yeast microarrays to learn the encoding system of yeast transcriptomic machinery. After model selection, we evaluated whether the trained models captured biologically sensible information. We show that the latent variables in the first hidden layer correctly captured the signals of yeast transcription factors (TFs), obtaining a close to one-to-one mapping between latent variables and TFs. We further show that genes regulated by latent variables at higher hidden layers are often involved in a common biological process, and the hierarchical relationships between latent variables conform to existing knowledge. Finally, we show that information captured by the latent variables provide more abstract and concise representations of each microarray, enabling the identification of better separated clusters in comparison to gene-based representation. CONCLUSIONS: Contemporary deep hierarchical latent variable models, such as the autoencoder, can be used to partially recover the organization of transcriptomic machinery.

Trans-species learning of cellular signaling systems with bimodal deep belief networks.

MOTIVATION: Model organisms play critical roles in biomedical research of human diseases and drug development. An imperative task is to translate information/knowledge acquired from model organisms to humans. In this study, we address a trans-species learning problem: predicting human cell responses to diverse stimuli, based on the responses of rat cells treated with the same stimuli. RESULTS: We hypothesized that rat and human cells share a common signal-encoding mechanism but employ different proteins to transmit signals, and we developed a bimodal deep belief network and a semi-restricted bimodal deep belief network to represent the common encoding mechanism and perform trans-species learning. These 'deep learning' models include hierarchically organized latent variables capable of capturing the statistical structures in the observed proteomic data in a distributed fashion. The results show that the models significantly outperform two current state-of-the-art classification algorithms. Our study demonstrated the potential of using deep hierarchical models to simulate cellular signaling systems. AVAILABILITY AND IMPLEMENTATION: The software is available at the following URL: http://pubreview.dbmi.pitt.edu/TransSpeciesDeepLearning/. The data are available through SBV IMPROVER website, https://www.sbvimprover.com/challenge-2/overview, upon publication of the report by the organizers. CONTACT: xinghua@pitt.edu SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.